Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
13
,
No.
1
,
Jan
uar
y
201
9
,
pp.
4
8
~
5
7
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v1
3
.i
1
.pp
4
8
-
5
7
48
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
Analysis
on swar
m robot
coordin
ation usi
ng fu
zzy logic
Ad
e
Sil
via
Handa
yani
1
,
Si
ti
Nu
rm
aini
2
, I
r
sy
adi
Yani
3
,
Nyayu
La
tif
ah
Husni
4
1,4
Depa
rtment
E
l
ec
tr
ic
a
l Engi
ne
er
ing
th
e
Pol
y
t
ec
h
nic
Sriwi
jay
a
,
Sr
iwij
a
y
a
Unive
rsit
y
.
Indon
esia
2
Facul
t
y
of
Com
pute
r
Sc
ie
nc
e, Sr
iwij
a
y
a
Unive
rsit
y
,
Indon
esia
3
Facul
t
y
of
Eng
i
nee
ring
,
Sriwi
jay
a
Univer
si
t
y
,
In
donesia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Sep
9
, 2
018
Re
vised N
ov 11
, 2
018
Accepte
d Nov
25, 201
8
In
thi
s
pape
r,
coor
dina
t
ion
among
indi
vidua
l
of
sw
arm
robot
in
comm
unic
at
ing
to
m
ai
ntain
th
e
safe
dist
ance
b
et
wee
n
robots
i
s
ana
l
y
z
ed
.
Ea
ch
robot
coo
rdina
t
es
the
ir
m
ovements
to
avoi
d
obstacles
and
m
oving
sim
ult
ane
ousl
y
.
Eva
lu
at
ion
of
sw
arm
robot
per
f
orm
anc
e
is
an
aly
z
ed
in
this
pape
r,
namel
y
:
t
he
coor
d
ina
t
ion
among
robots
to
share
information
in
sa
f
e
dista
nc
e
determ
ina
ti
on
.
In
cont
ro
ll
ing
th
e
coor
din
at
ion
of
m
oti
on,
ea
ch
robo
t
has
a
s
ensor
that
provid
es
sev
eral
input
s
about
i
t
s
surrounding
en
vironment.
Fuzz
y
logi
c
co
ntrol
in
thi
s
pa
per
al
lows
unc
ert
a
in
input,
an
d
produc
es
unli
m
it
ed
com
m
ands
to
cont
rol
m
oti
on
dir
ec
t
ion
with
spee
d
set
ti
ngs
ac
cor
d
ing
to
env
ironmenta
l
cond
it
ions.
In
thi
s
ex
per
iment,
it
is
o
bta
in
ed
that
the
si
ze of
th
e en
vironment
aff
ect
s the
coor
din
at
io
n
of
robots
.
Ke
yw
or
ds:
Com
m
un
ic
at
io
n
Coordi
nation
Fu
zzy
lo
gic
Sw
arm
r
ob
ot
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed.
Corres
pond
in
g
Aut
h
or
:
Sit
i Nurm
ai
ni
,
Faculty
of Com
pu
te
r
Scie
nc
e,
Sr
iwi
j
ay
a Univ
ersit
y
, Indo
nes
ia
.
Em
a
il
:
siti_n
ur
m
ai
ni@u
nsri.a
c.id
1.
INTROD
U
CTION
Coordi
nation
a
m
on
g
i
nd
i
vidual
r
obots
of
swar
m
is
on
e
of
inte
resti
ng
top
ic
s
in
rob
otic
sci
ence.
A
bette
r
c
omm
un
ic
at
io
n
am
on
g
them
becom
es
a
sig
nifica
nt
ne
ed
[
1]
.
Ea
ch
of
them
sh
oud
be
a
ble
to
tr
ansm
it
and
distri
bu
te
the
in
fo
rm
at
io
n
they
ha
ve
to
the
oth
e
r
r
obot
[2
-
3]
.
The
se
abili
ti
es
cou
ld
su
pp
or
t
t
he
robo
ts
t
o
detect
the
loc
at
ion
of
the
oth
e
r
r
obots,
to
sen
d
a
nd
t
o
recei
ve
in
f
or
m
at
ion
am
o
ng
t
hem
within
the
com
m
un
ic
at
ion
range,
so that
they can
p
e
rfo
rm
the task
c
ollec
ti
vely
.
Fo
r
i
ncr
ea
sin
g
the
perf
or
m
ance
of
c
omm
un
cat
ion
syst
em
in
swa
rm
ro
bots,
so
m
e
research
ers
ha
ve
pro
po
se
d
a
co
m
m
un
ic
at
ion
ne
twotk.
It
is
ve
ry
us
e
f
ul
f
or
im
pr
ov
in
g
s
wa
r
m
distribu
te
d
s
ensin
g
a
nd
detect
ing.
It
has
s
how
n
it
s
su
ccess
fu
l
in
var
i
ou
s
a
ppli
cat
ion
s,
s
uc
h
as:
form
ation
co
nt
ro
l
[4
-
5]
,
m
ulti
-
ta
rg
et
trac
king
[
6]
,
search
a
nd
resc
ue
[
7]
,
en
vir
onm
ental
m
on
i
toring
[
8]
,
an
d
surveil
la
nce
[9
-
10]
.
H
ow
e
ve
r
,
t
he
us
e
r
s
hould
kn
ow
wh
ic
h
c
omm
u
nicat
ion
they
c
an
us
e
i
n
thei
r
ap
plica
ti
on
.
T
his
pap
e
r
pr
es
ents
the
analy
s
is
on
c
omm
un
i
cat
ion
a
m
on
g t
he
i
ndivid
u of swa
rm
r
ob
ots in
c
onduct
ing co
ordi
na
ti
on
.
Since
com
m
un
ic
at
ion
netw
ork
de
vel
op
m
ent
has
beca
m
e
on
e
of
th
e
m
ai
n
chall
eng
e
s
in
s
war
m
rob
ots,
m
any
s
ign
ific
a
nt
deve
lop
m
ents
in
wireless
com
m
un
ic
at
ion
te
chnol
og
y
am
on
g
r
obots
has
bee
n
m
ade,
su
c
h
as:
NF
C,
W
i
-
Fi,
Bl
uetoo
t
h,
I
rDA,
G
SM
and
ZigB
ee
[11
–
14]
.
T
hese
te
ch
no
l
ogy
hav
e
e
nab
l
ed
the
dev
el
op
m
ent
of
a
utono
m
ous
ai
r,
gro
und,
or
unde
rwa
te
r
rob
ots.
N
FC
(N
ea
r
Fi
el
d
Com
m
un
ic
at
ion
)
and
Bl
uet
oo
t
h
are
consi
de
red
no
t
su
it
able
f
or
swar
m
ro
bots
du
e
to
lim
it
a
tio
ns
in
netw
ork
siz
e
[15]
.
The
W
i
-
Fi
-
base
d
a
pproach
m
ay
be
ad
op
te
d
f
or
a
sm
al
l
gr
oup
of
robo
ts
with
h
i
gh
perform
ance,
but
it
is
no
t
desirab
l
e
for
s
war
m
ro
bots
due
to
the
high
syst
em
com
plexit
y
and
hi
gh
c
os
t.
In
c
ontrary,
I
rDA
a
nd
ZigBee
a
re
widely
acce
pted
i
n
s
w
arm
ro
bots
bec
ause
of
the
l
ow
com
plexity
in
hard
war
e,
relat
ively
easy
syst
e
m
i
m
ple
m
entat
ion
,
an
d l
ow
powe
r
cons
um
ption
[
16
-
17]
. I
n
this
researc
h, A
Zigb
ee
co
m
m
un
ic
at
ion
was use
d.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An
alysis o
n
sw
ar
m
r
obot c
oo
r
dinati
on usi
ng
fuzzy lo
gic
(
Ad
e S
il
vi
a Han
dayan
i
)
49
In
pe
rfor
m
ing
the
com
m
un
ic
at
ing
ta
s
k,
t
he
rob
ot
m
us
t
be
able
to
c
oor
dina
te
on
e
ant
her
so
t
hat
they
can
a
void
c
olli
sion,
to
ke
ep
off
t
he
obsta
cl
es
an
d
t
o
m
antai
n
t
heir
distance
to
t
he
oth
e
r
robo
t
s
(
by
m
ai
nt
ai
ni
ng
an
accu
rate
spe
ed
to
the
nea
r
est
ro
bots)
.
Ma
intai
ning
the
distances
in
s
war
m
ro
bots
is
i
m
po
rtant
in
order
t
o
ob
ta
in
good
co
ordinati
on
to
a
chieve
co
ntr
olled
directi
on
[
18]
.
Fu
zzy
lo
gic
is
on
e
of
a
ppr
oac
hes
that
can
be
i
m
ple
m
ented
i
n
co
ntr
olli
ng
the
directi
on
of
swar
m
ro
bots.
it
has
been
su
c
cessf
ully
and
widely
us
e
d
to
con
t
rol
the
m
otion
of
s
war
m
robo
ts
[
19
]
.
This
te
ch
ni
qu
e
ca
n
sho
rten
the
ti
m
e
and
ref
ine
the
m
ov
e
m
ent
of
r
obot
s
in
a
ver
y
c
om
plex
syst
e
m
,
so
tha
t
it
can
av
oi
d
ob
sta
cl
es
[
13
-
14]
,
[
20
]
.
Fu
z
zy
log
ic
is
on
e
of
the
m
os
t
us
ef
ul
m
et
ho
ds
of
co
m
pu
ta
ti
on
al
intel
li
gen
ce
that
offer
s
t
he
ef
fi
ci
ency
an
d
sim
pl
ic
it
y
[21
–
23]
.
This
syst
e
m
s
us
e
li
ng
uisti
c term
s that are
sim
i
l
ar to t
hose th
at
hum
an
bein
gs
us
e
[24
-
25]
.
The
obj
ect
ive
of
t
his
pa
pe
r
is
to
eval
uate
th
e
perf
or
m
ance
of
com
m
un
ic
at
ion
am
on
g
m
ob
il
e
ro
bots
in
kee
ping
a
nd
coor
din
at
in
g
t
heir
m
otion
.
T
hey
sho
uld
be
able
to
m
ov
e
i
n
the
sam
e
dir
ect
ion
a
nd
m
ain
ta
in
their
pre
-
deter
m
ined
posit
io
ns.
T
o
achi
e
ve
coor
din
at
io
n
a
m
on
g
in
div
id
ua
l
of
s
wa
rm
ro
bots,
i
n
this
work,
a
wireless
c
omm
un
ic
at
ion
wa
s
us
e
d.
Each
in
div
id
ual
m
us
t
m
ai
ntain
a
pre
determ
ined
po
sit
ion
a
nd
ori
entat
io
n
a
m
on
g
them
w
hen
they
m
ov
e
in
t
heir
surr
ou
nd
i
ng.
H
ow
e
ve
r,
t
he
relat
ive
posit
ion
of
th
e
r
obots
is
not
fixe
d.
In
t
heir
fr
ee
m
ov
em
ent
of
ea
ch
r
obots
,
it
is
dif
ficult
to
know
a
suffici
e
nt
rob
ot
distance
to
obsta
cl
es
a
nd
t
o
oth
e
r
r
obots.
Using
fu
zzy
l
og
ic
as
the
s
war
m
arti
fici
al
intel
li
gen
ce
in
this
stu
dy,
m
ade
the
m
otion
coor
din
at
io
n
c
an be c
o
nt
ro
ll
e
d base
d on in
put dista
nce to g
ener
at
e c
orrect
d
eci
sio
n for t
he
outp
ut.
Fo
r
c
oor
din
at
i
on,
each
rob
ot
com
m
un
ic
ated
by
us
in
g
wireless
c
omm
un
ic
at
ion
,
X
-
Be
e
m
od
ule.
X
-
Be
e
m
od
ule
with
The
Re
c
ei
ved
Si
gn
al
S
tren
gth
I
ndic
at
or
(R
SSI)
as
a
par
am
et
er
to
est
i
m
at
e
the
distance
betwee
n
tw
o
X
-
Be
e
node
s.
I
n
this
wor
k,
X
-
B
ee
has
be
en
c
hose
n
as
a
wirel
ess
com
m
un
ic
at
ion
m
od
ule
a
m
on
g
rob
ots.
The
pur
pose
of
wir
el
ess
com
m
un
ic
at
ion
was
t
o
fin
d
the
r
obot
posit
io
n
in
the
ex
per
i
m
ental
env
i
ronm
ent. Th
e RS
SI i
ndic
at
or
is i
n
-
d
B
m
u
nits that
w
a
s u
se
d
t
o
m
easur
e
sig
nal stre
ngth
betwee
n r
obots
.
2.
SWAR
M RO
BOT
S COO
R
DINATIO
N
In
m
ai
ntaning
the
co
ordi
nation,
eac
h
rob
ot
in
the
swa
rm
m
us
t
hav
e
t
he
abili
ty
to
co
ordinate
a
nd
sh
are
the
wor
klo
a
d
to
t
he
ot
her
.
H
ow
e
ve
r
,
so
m
e
prob
le
m
s
a
lway
s
oc
c
ur
i
n
c
oord
i
na
ti
on
,
s
uc
h
as
duti
es
al
locat
ion
for
the
group
of
r
obots,
incl
ud
i
ng
:
res
ourc
es
us
a
ge;
ti
m
e
ta
sk
acc
ompli
sh
m
ent;
excessive
com
m
un
ic
at
ion
, s
e
nsor
selec
ti
on
s
, s
yst
em
r
el
ia
bili
ty
, an
d
sc
al
abili
ty
[26]
. Som
e researc
he
rs
trie
d t
o o
ve
rco
m
e
the
pro
blem
s
by
m
aking
som
e
i
m
pr
ov
em
ents
[
27
–
31]
.
Kam
ink
a
et
al
.
[27]
pro
pose
d
ef
fecti
ve
nes
s
ind
e
x
wh
ic
h
can
re
duce
tim
es
and
resou
rces
duri
ng
c
oord
i
natio
n
process
.
V.
Garg
[
28
]
desc
ribe
d
the
adva
ntages
of
us
in
g
r
obot
-
se
ns
or
netw
orks.
This
netw
ork
is
ver
y
us
ef
ul
for
co
ordi
natin
g
m
ulti
ple
ro
bots
or
s
war
m
ro
bots.
It
can
s
upport
the
s
war
m
to
s
har
e
sen
sor
dat
a
an
d
tra
c
k
it
s
m
e
m
ber
s.
T
o
e
nh
a
nce
the
li
fe
tim
es
of
netw
orks
,
A.
Wic
hm
ann
est
ablishe
d
t
he
sens
or
for
r
obot
c
omm
un
ic
at
ion
a
nd
c
oor
din
at
io
n
[
29]
that
can
re
duce
t
he
energy
us
a
ge.
Cork
eet
.al
[
30]
al
so
analy
z
ed
the
r
obot
that
wo
r
ke
d
tog
et
he
r
us
in
g
a
sens
or
ne
twork
.
M.
Schw
a
ge
r
in
[
31
]
us
e
d
se
ns
or
net
wor
k
of
so
m
e
no
des
.
These
se
ns
ors
hav
e
ca
pab
il
it
y
to
sense
the
va
lue
of
the h
i
gh se
ns
or
y functi
on
of
a
n
a
rea. It wil
l
de
te
ct
the obse
r
vation i
n hi
gh
e
r dens
it
y.
3.
SWAR
M RO
BOT
C
O
M
M
UN
I
C
ATIO
N DESIG
N A
N
D MET
HO
D
This
sect
io
n
e
xp
la
ine
d
t
he
r
esults
of
rese
a
rch
ga
ve
a
c
om
pr
ehen
sive
di
scussion
of
s
war
m
robo
t
com
m
un
ic
at
ion
.
The
Re
su
lt
s
are
pr
e
sente
d
in
fig
ur
es
,
graphs,
ta
bles,
a
nd
ot
her
s
to
m
ake
the
reade
r
easi
ly
unde
rstan
d
the
issues i
n
s
warm
r
ob
ot c
omm
un
ic
at
io
n.
3.
1
.
Desig
n
The
c
omm
un
ic
at
ion
m
od
el
in
colle
ct
ive
beh
a
vior
is
a
n
im
po
rtant
e
lem
ent.
It
rel
at
es
to
the
inf
or
m
at
ion
be
ing
distrib
uted
to
the
gro
up
[32]
.
T
her
e
a
r
e
a
lot
com
mu
nicat
io
n
m
od
el
s
of
g
r
oup
a
nim
a
ls
beh
a
vior
that
c
an
be
im
it
at
ed,
su
c
h
as
m
et
ric
[25
-
26]
,
the
to
po
l
og
ic
al
,
a
nd
visu
al
m
od
el
s
[
1]
,
[
33
]
.
The
m
et
ric
m
od
el
is
directl
y
based
on
s
patia
l
pr
oxim
i
t
y
wh
e
re
tw
o
ind
i
viduals
interact
if
they
are
within
a
cer
ta
in
distance
of
on
e
ano
the
r
[
32
]
,
[34]
.
To
po
l
ogic
al
m
od
el
need
s
each
r
ob
ot
to
interact
with
seve
ral
lim
it
ed
nu
m
ber
s
of
ne
arest
gro
up
m
e
m
ber
s
[35]
.
Th
e
visu
al
m
od
el
per
m
it
s
an
ind
i
vidual
to
inte
ra
ct
with
ot
her
a
gen
ts
in it
s v
is
ual f
ie
ld b
a
sed
on t
he
sen
s
or
y
capa
bi
li
t
ie
s o
f a
nim
a
ls [21].
To
dete
rm
ine
t
he
pe
rfor
m
anc
e
of
diff
e
ren
t
c
oor
din
at
io
n
of
colle
ct
ive
m
ove
m
ent
al
go
rith
m
s,
the
set
of
m
et
rics
is
use
d
that
ca
n
be
ap
plied
on
ly
f
or
f
or
m
at
i
on
s
or
f
or
floc
king
,
not
f
or
both
.
Du
e
to
t
he
di
fferent
natu
re,
th
ere
a
re
s
om
e
m
e
tri
cs
that
are
use
d
to
c
ha
racteri
ze
the
floc
k
ty
pe.
Dif
fer
e
nt
subsets
hav
e
be
e
n
determ
ined
by
div
idi
ng the
se
t of m
et
rics to g
r
oup t
he
s
ubs
et
accordin
g
t
o i
ts resem
blances
[
33
]
.
The
to
po
l
ogic
al
and
vis
ual
m
od
el
s
a
re
usual
ly
us
ed
f
or
perform
ing
the
m
et
ric
m
od
el
in
reachi
ng
t
he
ta
rg
et
[
34]
.
H
ow
e
ve
r,
the
re
is
no
cl
ea
r
dif
fer
e
nce
betwe
en
the
visu
al
and
t
opologica
l
m
od
el
s.
T
he
visu
al
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
4
8
–
5
7
50
m
od
el
latency
was
substant
ia
ll
y
lower
th
an
the
topolo
gical
and
m
etr
ic
m
od
el
s;
bu
t,
the
m
e
tric
m
od
el
ou
t
perform
ed
the to
polo
gical
m
od
el
in
te
rm
s
of the
transfe
r of i
nfo
rm
ation
[35].
Sele
ct
ing
a
to
po
l
og
y
that
sui
ta
ble
to
ou
r
s
war
m
ro
bot
char
act
erist
ic
ne
eds
is
an
im
portant
ta
sk.
In
pa
rtic
ular,
how
t
o
f
orm
and
m
ai
ntain
an
unbro
ken
com
m
un
ic
at
ion
netw
ork
dynam
ic
ally
so
that
the
inf
or
m
at
ion
th
rou
gh
the
swa
rm
can
r
un
co
ntinu
ously
f
or
the
entire
s
warm
beco
m
es
an
interest
in
g
prob
le
m
[36]
.
RS
SI
is
one
of
the
s
olu
ti
on
s
in
this
pro
blem
.
It
is
us
e
d
as
a
c
om
plem
entary
too
l
to
co
ns
ide
r
t
he
t
opology
of
t
he
entire
l
ocali
zat
ion
sys
tem
[37]
.
I
n
gen
e
ral,
the
s
war
m
ro
bot
pe
rfor
m
ance
is
aff
ect
e
d
by
ne
twork
topolo
gy
on
no
ise
est
im
at
i
on
a
nd
r
obust
ness.
T
opol
ogy
of
the
network
co
uld
pract
ic
al
ly
aff
ect
the
perform
ance o
f
algorit
hm
s f
or large
interc
on
nected
swa
rm
r
obots syste
m
[38]
.
Bl
ock
diag
ram
co
ntro
l
of
co
m
m
un
ic
at
ion
a
m
on
g
swa
rm
ro
bots
is
pr
ese
nt
ed
in
Fi
gure
1.
The
X
-
Be
e
or
Zig
bee
prot
oco
l
t
hat
is
c
onnecte
d
to
the
central
c
om
pu
te
r
c
ollec
ts
exp
e
rim
ental
data
us
e
d
as
a
us
ef
ul
com
m
un
ic
at
ion
to
cont
r
ol all
existi
ng syst
em
s o
n
the act
ua
l rob
ot p
la
tf
orm
.
Figure
1. Bl
oc
k diag
ram
co
ntro
l
of swa
rm
r
obot c
oor
din
at
ion
X
-
Be
e
or
Zi
gB
ee
protoc
ol
ba
sed
m
od
em
s
su
pp
or
t
th
ree
diff
e
re
nt
netw
ork
to
po
l
ogie
s
ie
sta
r,
m
esh,
and
cl
us
te
r
t
re
e
netw
orks
,
al
lowing
a
va
riet
y
of
cu
stom
ized
c
onfig
ur
at
i
on
s
.
Co
ordi
nator,
a
set
of
routers
,
and
e
nd
de
vice
s
are
com
m
on
things
that
topolo
gy
m
esh
pocesses
[
10
]
.
A
router
ca
n
be
li
nk
e
d
to
one
or
m
or
e
routers
a
nd
en
d
de
vices.
T
he
com
m
un
ic
at
ion
r
ules
of
m
es
h
to
po
l
og
y
a
re
flexib
le
be
ca
use
the
r
ou
te
rs
that
are
locat
ed
within
the
ra
nge
of
e
a
ch
oth
e
r
ca
n
c
omm
un
ic
at
e
directl
y.
A
n
a
dvantage
of
the
m
esh
net
work
is
that
there
is
od
ds
-
on
a
nothe
r
al
te
rn
at
ive
route
in
case
an
e
xisti
ng
li
nk
fa
il
s.
Ther
e
by,
this
ty
pe
of
ne
twor
k
topolo
gy is c
onsis
te
ntly
good
in qu
al
it
y o
r p
erfor
m
ance.
In
pa
pe
r
[
39
]
,
it
was
exp
la
in
ed
the
us
e
of
RSSI
in
trac
kin
g
the
s
war
m
rob
ot.
The
co
m
m
un
ic
at
ion
us
e
d
is
cente
re
d
w
he
re
the
r
obot
becam
e
a
l
eader
a
nd
the
f
ollow
e
rs
c
omm
un
ic
at
e
wirelessl
y
throu
gh
X
-
Be
e.
The
dif
fer
e
nce
of
that
resea
rc
h
with
this
res
earch
is
the
r
obot
strat
egy
i
n
coor
din
at
in
g
it
s
own
m
ov
em
ents
to
avo
i
d
the
obsta
cl
es an
d
c
olli
si
on to othe
r rob
ots. In t
his
rese
arch, a
fu
zzy
l
og
ic
was
u
se
d.
3.2
.
F
uz
z
y
L
og
ic
f
or Coor
dinat
i
on
In
c
oor
din
at
in
g
the
s
war
m
ro
bot,
eac
h
in
di
vid
ual
rob
ot
s
war
m
m
us
t
hav
e
abili
ty
to
coope
rate
to
perform
a
sp
eci
fic
ta
sk
,
as
well
as
the
robo
t
m
us
t
be
a
ble
to
interact
with
the
env
i
ronm
ent.
The
work
i
ng
env
i
ronm
ent
of
swa
rm
ro
bo
t i
s
com
plex
and
changea
ble;
in
add
it
io
n
each r
ob
ot
con
sist
s o
f
m
any
co
m
po
ne
nts
su
c
h
as
c
omm
un
ic
at
io
n
de
vices,
syst
em
con
tr
ol,
se
ns
in
g
et
c.,
m
aking
it
dif
ficult
to
de
te
rm
ine
m
a
them
at
ic
al
m
od
el
s.
It
is
qu
it
e
im
po
ssib
le
to
identify
.
Fu
zzy
log
ic
al
gorithm
of
fer
s
the
so
l
utio
ns
by
ig
norin
g
the
m
at
he
m
at
ic
a
l equ
at
io
ns.
Fu
zzy
Lo
gic
The
or
y
is
a
decisi
on
-
m
akin
g
te
chn
i
qu
e
th
at
translat
es
values
ex
pr
esse
d
in
la
ngua
ge
(linguist
ic
s)
in
to
spe
ci
fic
va
lue
s,
w
hich
m
ay
be
diff
ic
ult
to
res
olv
e
with
tra
diti
on
a
l
m
at
he
m
at
ic
s
[34]
.
The fuzzy l
og
i
c co
ns
ist
ing o
f l
ing
uisti
c c
on
t
r
ol rules that
is
desig
ne
d
as c
oor
din
at
e m
otion
c
on
t
ro
ll
er
ba
sed o
n
the
knowle
dge
and
e
xperie
nc
e
of
t
he
hum
a
n
ex
pe
rt
[35
-
36]
.
I
n
the
m
ove
m
ents,
co
ordinati
on
c
ontr
oller
will
X
b
e
e
P
e
r
s
o
n
a
l
C
o
m
p
u
t
e
r
I
n
t
e
r
f
a
c
e
C
o
o
r
d
i
n
a
t
e
d
S
y
s
t
e
m
C
o
m
m
a
n
d
S
e
n
s
o
r
1
S
e
n
s
o
r
2
S
e
n
s
o
r
3
A
r
d
u
i
n
o
l
e
f
t
r
i
g
h
t
P
r
o
c
e
s
s
i
n
g
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
m
o
t
o
r
X
b
e
e
R
o
b
o
t
1
C
o
m
p
a
s
s
S
e
n
s
o
r
1
S
e
n
s
o
r
2
S
e
n
s
o
r
3
A
r
d
u
i
n
o
l
e
f
t
r
i
g
h
t
P
r
o
c
e
s
s
i
n
g
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
m
o
t
o
r
X
b
e
e
R
o
b
o
t
2
C
o
m
p
a
s
s
S
e
n
s
o
r
1
S
e
n
s
o
r
2
S
e
n
s
o
r
3
A
r
d
u
i
n
o
l
e
f
t
r
i
g
h
t
P
r
o
c
e
s
s
i
n
g
a
n
d
C
o
m
m
u
n
i
c
a
t
i
o
n
m
o
t
o
r
X
b
e
e
R
o
b
o
t
3
C
o
m
p
a
s
s
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An
alysis o
n
sw
ar
m
r
obot c
oo
r
dinati
on usi
ng
fuzzy lo
gic
(
Ad
e S
il
vi
a Han
dayan
i
)
51
tur
n
the
r
obot
wh
eel
with
a
c
on
sta
nt
ra
ng
e
t
hro
ugh
the
fu
z
zy
con
tr
oller.
The
co
ndit
ion
s
us
ed
i
n
the
c
ontr
oller
dep
e
nd
on the
m
ov
e
m
ent o
f
t
he ro
bot.
In
this
researc
h,
3
swa
rm
ro
bo
ts
wer
e
us
e
d.
Eac
h
r
obot
sh
oul
d
achiev
e
the
ta
sk
of
m
ov
i
ng
t
o
the
destinat
io
n
a
nd
a
vo
i
d
obsta
cl
es.
Th
us,
e
ve
ry
r
obot
in
t
he
s
war
m
has
three
ta
sk
s:
avo
i
ding
obsta
cl
es,
m
ov
ing
to
th
e
destinat
io
n,
as
well
as
keep
i
ng
t
he
swar
m
by
avo
i
ding
colli
sion
s
am
ong
robo
ts
.
The
e
nv
ir
onm
ent
us
e
d
was
an
en
vir
onm
e
nt
without
obst
acl
es
with
di
fferent
ar
ena
s
ha
pes
a
nd
siz
es
.
I
n
an
env
i
ronm
ent
without
obsta
c
le
s,
there
was
no
distu
rbanc
e
eff
ect
s
occured
.
I
f
the
rob
ot
was
far
fro
m
that
gro
up,
then
th
e
robo
t
w
ould
m
ov
e
towa
rds
on
e
a
no
t
her
to
def
e
nd
the
swar
m
.
If
eac
h
r
obo
t
was
c
losed
,
the ro
bo
t
ha
d
t
o
sta
y a
way on
e an
oth
e
r
to
avoid a
co
ll
isi
on.
In
c
oord
i
natio
n
of
the
swa
r
m
ro
bot,
the
interact
io
n
bet
ween
rob
ots
in
the
swa
rm
d
epends
on
the
distance
betwe
en
the
r
obots
with
the
obsta
cl
es
an
d
with
oth
e
r
r
obots
de
te
ct
ed
from
ea
ch
se
nsor
.
By
us
i
ng
fu
zzy
lo
gic,
th
e
sensor
in
pu
t
s
of
eac
h
r
obot
are
the
input
value
f
or
the
m
e
m
ber
sh
ip
f
unct
ion
(MF)
.
I
n
this
researc
h,
f
uzz
y
con
t
ro
l
desi
gn
is
s
how
n
i
n
Fi
g
ure
2.
T
he
fu
zzy
co
ntr
ol
str
uctu
re
ba
sed
on
th
e
pro
cedure
consi
sts
of
th
e
sta
ndar
d
procedu
res,
s
uc
h
as:
in
put
c
rips,
f
uzzifica
ti
on
,
f
uzzy
in
pu
t,
r
ule
ev
al
uatio
n,
fu
zzy
outp
ut, d
efu
zzi
ficat
io
n and o
utput c
rip
s
Figure
2. Desi
gn
f
uzzy lo
gic
con
t
ro
ll
er
Ba
sed
on
the
va
lue
of
M
F,
s
om
e
ru
le
s
for
t
he
res
pons
e
of
t
he
m
oto
r
outp
ut
of
t
he
s
warm
ro
bo
t
will
be
m
ade.
In
thi
s
stud
y,
the
i
nput
a
nd
ou
t
pu
t
values
of
MF
are
sho
wn
in
F
ig
ure
3
(a
)
a
nd
(b).
I
n
Fig
ure
3
(a
),
there a
re tw
o m
e
m
ber
sh
ip
f
unct
ions (MF
s),
i.e. f
a
r
a
nd n
ea
r.
B
oth
of them
are
in
the t
rapez
oid
al
form
o
f
MFs
.
In Fig
ure
3 (
b)
t
he
(c
onseq
ue
nt
)
outp
ut
of
t
he
syst
e
m
is n
ot
a f
uzzy set
, but
r
at
he
r
tha
n a c
on
sta
nt
or
a
li
ne
ar.
(a)
Me
m
ber
s
hip
functi
on
of
di
sta
nce senso
r a
s
input
(b)
Me
m
ber
sh
i
p
functi
on
of
m
oto
r
sp
ee
d
as
ou
t
pu
t
Figure
3. I
nput
MF and
outp
ut
MFvalue
The
ru
le
set
in
the
fu
zzi
ficat
i
on
pr
ocess
i
n
t
he
form
of
c
ontrol
decisi
ons,
resu
lt
in
g
in
a
c
om
bin
at
ion
of
in
pu
t
a
nd
outp
ut.
In
this
st
ud
y
8
r
ules
we
re
us
ed
as
s
ho
wn
in
Table
1.
It
pr
ese
nts
li
nguisti
c
var
ia
ble
as
th
e
ou
t
pu
t c
ontr
oller
wh
ic
h
c
on
ta
ins the
m
oto
r
s
peed
f
or the
rig
ht and le
ft P
W
M.
F
u
z
z
i
f
i
e
r
I
n
f
e
r
e
n
c
e
D
e
f
u
z
z
i
f
i
e
r
R
u
l
e
B
a
s
e
I
n
p
u
t
C
r
i
p
s
O
u
t
p
u
t
C
r
i
p
s
F
u
z
z
y
L
o
g
i
c
C
o
n
t
r
o
l
l
e
r
O
b
s
t
a
c
l
e
R
a
n
g
e
S
p
e
e
d
a
n
d
S
t
e
e
r
i
n
g
C
o
n
t
r
o
l
D
i
s
t
a
n
c
e
S
e
n
s
o
r
(
L
e
f
t
,
F
r
o
n
t
,
R
i
g
h
t
)
M
o
t
o
r
V
e
l
o
c
i
t
y
(
L
e
f
t
a
n
d
R
i
g
h
t
P
W
M
)
2
0
4
0
6
0
8
0
1
0
0
1
1
0
3
0
5
0
7
0
9
0
n
e
a
r
f
a
r
1
s
l
o
w
4
0
m
e
d
i
u
m
1
2
0
f
a
s
t
2
5
0
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
4
8
–
5
7
52
Table
1.
Fu
zzy
Logic Rule
Based
Ru
les
Distan
ce Sens
o
r
as
I
n
p
u
t
Moto
r
Sp
eed as O
u
tp
u
t
Left
Fo
rwar
d
Rig
h
t
Left
PWM
Rig
h
t PW
M
1
Near
Near
Near
Slo
w
Fast
2
Near
Near
Far
Fast
Slo
w
3
Near
Far
Near
Mediu
m
Mediu
m
4
Near
Far
Far
Mediu
m
Slo
w
5
Far
Near
Near
Slo
w
Fast
6
Far
Near
Far
Slo
w
Mediu
m
7
Far
Far
Near
Slo
w
Mediu
m
8
Far
Far
Far
Fast
Fast
4.
RESU
LT
S
AND A
N
ALYSIS
In
t
his
re
searc
h,
t
hr
ee
r
obots
wer
e
util
iz
ed.
The
real
desi
gn
of
the
r
obot
s
are
s
how
n
i
n
F
i
gure
5.
Ever
y
r
obot
ha
s
three
distanc
e
sensors,
one
com
pass
sensor,
an
d
one
X
-
B
ee.
The
r
obots
with
ci
rcu
la
r
s
hape
hav
e
diam
et
er
15
cm
and
he
igh
t
17
cm
.
The
rob
ot
us
e
s
three
w
heels
,
tw
o
rea
r
wheel
s
of
t
he
r
obot
ha
ve
functi
oned
as
a
co
ntr
oller,
one
wh
e
el
has
f
un
ct
io
ne
d
as
f
r
eel
y
m
ov
er.
T
wo
DC
m
oto
rs
are
c
onne
ct
ed
to
t
he
two
dri
vi
ng
w
he
el
s
resp
ect
ive
ly
.
The
ro
ta
ti
on
directi
on o
f
e
ach
m
oto
r
was
con
t
ro
ll
ed
by
the
directi
on o
f
dr
i
ve
current,
whil
e t
he rotat
io
n
s
pe
ed was c
ontroll
ed by t
he d
uty cy
cl
e o
f Pulse
W
i
dth M
od
ulati
on
(P
WM)
.
This
e
xp
e
rim
e
nt
was
do
ne
in
an
in
door
e
nv
iro
nm
ent.
The
te
st
aren
a
us
e
d
is
1×
6
m
,
2×4
m
and
3×
4
m
as
sh
own
in
Figure
4.
T
he
rob
ot
m
ov
ed
al
ong
the
pr
e
s
et
path
within
the
sco
pe
wh
il
e
m
ai
ntaining
i
ts
own
po
sit
io
ns
.
T
he
y
m
ov
ed
al
on
g
the
f
our
si
de
s
of
a
s
qu
are
.
Each
rob
ot
ha
d
a
n
e
qu
i
valen
t
beh
a
vior
a
nd
sam
e
local
iz
at
ion
process.
1×6 m
2×4
m
3
×4m
Figure
4. Ex
pe
rim
ental
en
vironm
ent an
d
t
he l
ocali
zat
ion
sy
stem
4.1
.
E
xp
eri
m
ent
al R
es
ults
of Swarm
Ro
bo
t
Co
ordin
ati
on
In
this
w
ork
we
pr
es
ent
t
w
o
kinds
of
e
xperim
ents:
(i)
co
ordinati
on
betwee
n
rob
ot
to
perf
or
m
colle
ct
ive
an
d
si
m
il
ar
directi
on
m
ov
em
ent.
Each
rob
ot
m
us
t
defen
d
pr
e
-
determ
ined
posit
ion
s
an
d
ori
enta
ti
ons
a
m
on
g
t
hem
a
t
the
sa
m
e
t
i
m
e;
(ii)
com
m
u
nicat
ion
betwe
en
swa
rm
s
of
rob
ots
to
exc
ha
ng
e
t
he
i
nform
at
ion
about the
sett
in
gs
of the m
otion
directi
on.
Coordi
nation
of
r
obots
can
be
en
han
ce
d
t
hro
ugh
c
omm
un
ic
at
ion
,
f
or
i
ns
ta
nce,
t
he
a
bili
ty
f
or
s
ensi
ng
ano
t
her
r
obot
.
The
co
ordin
at
ion
am
on
g
the
rob
ots
reli
es
on
net
wor
k
com
m
un
ic
at
ion
.
In
te
rm
of
it
s
netw
orkin
g
ca
pab
il
it
y,
ever
y
rob
ot
com
m
u
nicat
ed
to
one
ano
t
her
only
at
even
t
ti
m
es
and
co
ordinat
ed
th
e
m
ov
ing
of
swa
rm
ro
bot
tr
ough
lo
w
-
po
wer
e
d
ra
dio
X
-
bee
.
Each
rob
ot
at
tem
pted
to
f
ollow
the
t
race
of
oth
e
r
rob
ots
by
sen
sing
their
sig
nal
stren
gt
hs
.
O
nc
e
they
r
eache
d
th
e
e
nd
of
t
he
trace,
they
w
il
l
travel
f
ur
t
he
r
int
o
the un
known
e
nv
i
ronm
ent u
nt
il
they can m
ain
ta
in a m
ini
m
a
l connect
io
n
t
o t
he rest
of t
he gr
oup.
In
this
researc
h,
the
i
nteracti
on
betwee
n
r
obots
to
c
oor
di
nate
de
pende
d
on
th
e
distanc
e
a
m
on
g
t
he
rob
ots
to
the
obsta
cl
es
an
d
to
oth
e
r
robo
ts
detect
ed
from
each
sens
or
s.
Usi
ng
the
f
uzzy
m
et
hod,
the
m
agn
it
ude
of
the
P
WM
m
otor
sp
ee
d
was
cal
culat
ed
us
ing
f
uz
zy
co
ntr
oller
base
d
on
the
m
agn
it
ude
of
th
e
per
cei
ved
distances.
F
uzzy
co
ntr
ollers
ha
d
th
ree
input
an
d
two
ou
t
pu
ts
tha
t
reg
ulate
d
t
he
rig
ht
and
le
ft
P
WM
sp
ee
ds
.
Thr
ee
a
ren
as
of
indo
or
ex
pe
rim
ents
scenario,1
x6
m
,
2×4
m
and
3x
4
m
with
the
obsta
cl
e
wer
e
al
so
cond
ucted
in
t
his
resea
rch.
T
he
est
i
m
at
ed
po
sit
ion
s
wer
e
r
el
at
ively
near
to
each
r
ob
ot
duri
ng
the
proc
ess
of
m
ov
ing
al
ong
the
first
are
na
of
the
2x4
m
.
Ho
we
ve
r,
the
r
obots
ha
d
di
ff
e
r
ent
directi
ons
and
diff
e
re
nt
relat
ive
distances t
o
ea
ch othe
r wh
e
n t
he
r
ob
ots m
ove f
ur
the
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An
alysis o
n
sw
ar
m
r
obot c
oo
r
dinati
on usi
ng
fuzzy lo
gic
(
Ad
e S
il
vi
a Han
dayan
i
)
53
4.1.1
1×
6
m
A
rena
Figure
5
s
how
s
that
thei
r
de
sti
nation
has
r
eached
by
al
l
ind
ivi
du
al
r
obots
a
nd
m
utu
al
ly
avo
id
e
d
colli
sion
am
ong
the
r
obots
a
long
the
way.
In
0
-
15
seco
nds,
the
s
wam
rob
ot
m
ov
ed
i
n
ta
ndem
.
In
the
40
seco
nd
s
,
the
r
obot
p
osi
ti
on
wer
e
cl
os
e
d
t
o
on
e
an
ot
her,
ho
wev
e
r
t
he
r
obot
did
no
t
colli
de,
m
oreov
e
r,
they
cou
l
d
av
oid
to
hit
the
wall
.
The
m
ove
m
ent
of
r
obot
s
wh
e
n
av
oidi
ng
ob
sta
cl
es
cou
l
d
be
see
n
in
the
screen
shot
of
r
eal
vid
e
o
im
age
as
sho
wn
in
Figure
4
(b).
F
or
first
10
se
c
onds,
the
r
obots
wer
e
dis
perse
d
a
nd
at
n
ext
50 sec
onds, t
heir posi
ti
on
s
w
e
re clo
s
ed
to
gethe
r
a
nd
recon
nected.
t
=
1
5
s
t
=
2
5
s
t
=
3
5
s
t
=
4
0
s
t
=
4
5
s
t
=5
0
s
t
=6
0
s
t
=6
5
s
t
=7
0
s
t
=8
0
s
t
=1
1
0
s
t
=1
2
0
s
t
=1
5
0
s
t
=2
0
0
s
(a)
(b)
Figure
5. S
warm
ro
bo
t c
oor
din
at
ion ex
pe
rim
ents in
are
na 1
x4 m
(
a) trac
kin
g r
obot
(b)
,
e
xp
e
rim
ent
photog
raphs
In
the
1
×
6
m
aren
a
,
the
m
ot
or
m
ov
ed
slo
w
er
than
the
ot
he
r
aren
a
.
This
is
becau
se
of
th
e
aren
a
ha
d
a
width
of
onl
y
10
0
cm
,
wh
il
e
the
di
m
ension
s
of
eac
h
robo
t
was
17
cm
.
The
r
obots
w
ou
l
d
searc
h
th
ei
r
safe
po
sit
io
ns
to
a
vo
i
d
obsta
cl
es.
In
this
a
ren
a
,
the
m
ov
e
m
ents
of
s
war
m
ro
bots
in
m
ai
nt
ai
nin
g
posit
io
n
an
d
directi
on
with
a certai
n dist
an
ce we
re
diff
ic
ul
t t
o
be
c
oor
dina
t
ed
due t
o
t
he nar
row
s
pace
of the
are
na.
4.1.2
2×
4
m
Are
na
In
are
na
of
2x
4
m
exp
e
rim
en
t,
the
e
nv
i
ron
m
ent
was
set
without
obsta
cl
es
as
in
a
re
na
1x6
m
before
.
In
Fig
ur
e
6(
a
),
It
can
be
see
n
the
grap
h
of
di
recti
on
a
ng
le
vs
ti
m
e
per
second.
It
can
be
fou
nd
t
hat
with
th
e
increasin
g
of
t
i
m
e,
the
directi
on
a
ng
le
gra
dual
l
y
reached
it
s
final
po
sit
ion.
Fig
ur
e
6(b
)
is
the
exp
e
ri
m
ent
photog
raphs
of
t=
5s
,
t=
15
s,
t
=25
s
,
t=
35s,
t=
50
s
a
nd
t=
60s,
resp
ect
ively
.
I
t
can
be
seen
that
the
robo
ts
m
ov
ed
in the sa
m
e d
irect
ion
a
nd the
sam
e d
ist
ances.
Fr
om
the
experim
ent,
it
sh
ow
e
d
that
th
r
ee
robo
ts
m
ov
ed
i
n
the
c
oor
din
at
io
n
one
an
oth
e
r.
Sw
arm
robo
ts
m
ov
e
m
ent
can
be
seen
i
n
Fig
ur
e
6.
All
the
r
obots
co
ul
d
av
oid
ob
sta
cl
es
a
nd
m
ov
e
d
ar
ou
nd
t
he
aren
a
in
sm
oo
t
h
m
ov
em
ent.
T
he
data
of
dire
ct
ion
an
gle
vs
t
i
m
e
are
sh
own
in
Figure
6.
From
the
gr
a
ph,
it
can
be
co
nclu
ded
t
hat
there
we
re
directi
on
an
gle
changes
vs
ti
m
e
per
second
in
the
ran
ge
of
10
-
15,
20
-
25,
45
-
50
and
75
-
80.
It
was
due
to
the
ro
bots
detect
e
d
the
obsta
cl
e
and
the
wall
.
Af
te
r
r
eachi
ng
a
safe
po
sit
io
n,
each
rob
ot
will
lower
the
sp
ee
d
and
w
ai
te
d
f
or
ano
t
her
r
obot
to
m
ov
e
bac
k
in
the
sam
e
directi
on
to
reac
h
the
sp
eci
fied
posit
ion.
On
ce
in
divi
du
al
rob
ots g
a
there
d
in t
he
s
pe
ci
fied desti
nat
ion
,
they
woul
d on
ly
m
ov
e ar
ound
the d
e
sti
nation
area.
1
2
3
1
3
3
1
2
2
1
2
3
t
=
6
0
s
t
=
4
0
s
1
3
2
t
=
1
5
s
t
=
2
0
0
s
t
=
8
0
s
6
0
0
c
m
1
0
0
c
m
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
4
8
–
5
7
54
t = 5 s
t = 1
5
s
t = 2
5
s
t = 3
5
s
t = 5
0
s
t = 6
0
(a)
(b)
Figure
6. S
warm
ro
bo
t c
oor
din
at
ion ex
pe
rim
ents in
are
na 2
x4 m
(
a) trac
kin
g r
obot
(b)
e
xperim
ent
photog
raphs
4.1.3
3
×
4
m
A
rena
In
a
s
ubseq
ue
nt
ex
per
im
ent,
to
show
t
hat
the
rob
ot
can
be
ap
plied
in
an
un
known
e
nv
i
ronm
ent,
so
m
e
ob
sta
cl
es
(
wall
s)
wer
e
gi
ven
in
t
he
a
re
na.
I
n
t
he
3×4
m
aren
a
in
Fig
ur
e
7
,
it
s
howe
d
t
hat
al
l
in
divi
du
al
rob
ots
had
m
ov
ed
po
sit
io
ns
a
long
with
one
ano
t
her
an
d
av
oid
e
d
obsta
cl
es
al
ong
the
wa
y.
At
t
he
25
se
cond
s
,
each
in
div
i
du
a
l
robo
t
a
ssem
bled
i
n
a
n
a
djacent
posit
io
n,
howe
ver,
the
y
cou
l
d
no
t
m
ov
e
fr
eel
y.
This
i
s
because
the
re
wer
e
obsta
cl
es
in
the
f
or
m
of
wall
s,
s
o
the
y
disp
er
sed
,
a
nd
a
fter
a
sa
fe
posit
ion
,
t
he
y
would
return
bac
k
t
oget
he
r
(at
45
s
econds
).
I
n
thi
s
en
vir
onm
ent,
the
m
ov
em
ents
of
t
he
s
war
m
r
obots
wer
e
a
ble
to
coor
din
at
e in
m
ai
ntaining
t
he
posit
ion an
d direct
io
n wit
h a certa
in
distan
ce.
Fr
om
severa
l
e
xp
e
rim
ents
that
had
bee
n
c
on
du
ct
e
d
in
a
dif
fer
e
nt
are
na,
it
co
uld
be
c
onc
lud
e
d
that
the
real
rob
ot
swar
m
m
ov
e
m
ent
was
the
in
div
id
ual
m
ob
il
e
robo
t
that
cou
l
d
achie
ve
it
s
go
al
s
effe
ct
ively
.
Mov
em
ent
of
the
swa
rm
ro
bot
co
-
ordi
natio
n
sho
wed
t
he
perform
ance
of
be
hav
i
or
i
n
searchi
ng
pur
pose.
At
the
m
o
m
ent
of
t
he
ad
j
ace
nt
posit
ion
,
the
ro
bot
w
ould
r
edu
ce
the
s
pee
d
an
d
r
otate
in
the
oth
er
dire
ct
ion
s
un
ti
l
it
reac
hed
the
sa
fe
posit
ion.
On
ce
the
s
afe
po
sit
io
n
got,
of
the
r
obot
would
gat
her
to
the
sam
e
locat
ion
and w
e
nt
hand i
n
ha
nd in
the
s
a
m
e d
irect
ion
.
t
=
5
s
t
=
2
5
s
t
=
60
s
t
=
1
0
0
s
t
=
1
4
0
s
t
=
1
5
s
t
=
4
5
s
t
=
80
s
t
=
1
2
0
s
t
=
1
6
0
s
(a)
(b)
Figure
7. S
warm
ro
bo
t c
oor
din
at
ion ex
pe
rim
ents in
are
na 3
x4 m
(
a) trac
kin
g r
obot(
b) e
xperim
ent
photog
raphs
1
2
3
t
=
0
s
1
2
3
3
3
1
2
1
2
3
1
2
t
=
1
5
s
t
=
2
5
s
t
=
6
0
s
4
0
0
c
m
t
=
3
5
s
2
0
0
c
m
1
2
3
1
2
3
1
2
3
1
2
3
1
3
2
3
0
0
c
m
t
=
4
5
s
1
2
3
4
0
0
c
m
t
=
2
5
s
t
=
6
0
s
t
=
1
6
0
s
t
=
1
0
0
s
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An
alysis o
n
sw
ar
m
r
obot c
oo
r
dinati
on usi
ng
fuzzy lo
gic
(
Ad
e S
il
vi
a Han
dayan
i
)
55
In
a
ren
a
1x6
m
,
the
m
oto
r
s
peed
m
ov
em
e
nt
was
slo
wer
than
oth
er
a
re
nas.
T
his
is
be
cause
of
the
aren
a
ha
d
the
width
of
10
0c
m
,
wh
il
e
the
di
m
ension
of
ea
ch
r
obot
17
cm
.
The
r
obot
would
decr
ease
s
peed
t
o
fin
d
a
sec
ur
e
posit
ion
i
n
a
vo
i
ding
ob
sta
cl
es.
Wh
il
e
on
the
l
arg
e
r
are
na
(2x
4
m
),
the
r
obot
m
ov
ed
in
t
he
sam
e
directi
on
an
d
c
oex
ist
e
d
al
m
os
t
ever
y
tim
e.
The
m
ov
em
ent
of
s
war
m
ro
bo
t
was
m
or
e
sta
ble.
I
n
the
a
re
na
of
3x4
m
with
the
lim
i
ti
ng
wall
s
ect
ion
,
t
he
tra
ve
l
tim
e
of
the
r
obot
was
sl
ower
tha
n
in
the
a
ren
a
2x4
m
.
This
is
because
of
the
p
os
it
io
n
of
th
e
adj
ace
nt
r
ob
ot
from
the
beg
in
ning
an
d
ea
ch
r
obots
slo
w
ed
do
wn
t
he
s
peed
t
o
achieve
a sec
ure p
os
it
ion.
4.2
.
Rssi
Me
as
ureme
nt
Re
cei
ved
Sig
na
l
Stren
gth
I
ndic
at
or
/R
SSI
is
the
sig
nal
le
ve
l
(in
-
dBm
)
of
la
st
good
pac
ke
t
receive
d.
Ther
e
are
t
wo
te
chn
iq
ues
t
o
r
ead
RSS
I
valu
e:
1)
RS
SI
val
ue
is
e
ncode
d
into
P
ulse
-
Wid
th
Mo
du
la
te
d
sign
al
avail
able
at
the
X
-
Be
e
m
odule,
an
d
2)
R
SSI
value
is
rea
d
via
an
API
c
om
m
and
.
Th
e
R
SSI
value
re
ported
by
X
-
Be
e
Prom
od
ule
is
bet
wee
n
-
36
to
-
100
dB
m
wh
il
e
that
of
a
sta
nd
a
rd
X
-
bee
m
od
ule
is
betwee
n
-
23
t
o
-
92
dBm
.
Ho
we
ve
r,
the
XBee
m
anu
al
say
s
that
the
re
port
ed
val
ue
is
accurate
be
twe
en
-
40
dBm
and
t
he
sensiti
vity
o
f
Xb
ee
m
od
ule’s
r
ecei
ve
r [25]
.
In
this
e
xperi
m
ent,
f
or
RSS
I
m
easur
em
ent
,
tw
o
XBee
P
ro
m
od
ules
(for
exam
ple,
one
node
is
a
Coordi
nato
r
an
d
the
oth
e
r
is
a
Rou
te
r/En
d
dev
ic
e
)
we
re
c
onnected
an
d
t
hen
t
he
distan
ce
betwee
n
the
m
was
var
ie
d
to
m
ea
su
re
t
he
relat
ion
s
hi
p
betwee
n
RSSI
values
and
distances
.
In
th
e
RSSI
read
i
ng
e
xp
e
ri
m
ents,
three
m
ob
il
e
ro
bots
co
ntaini
ng
t
he
X
-
Be
e
Series
m
od
ul
es,
one
as
Co
ordinat
or
a
nd
oth
e
r
as
Ro
uter/En
d
dev
ic
e.
This expe
rim
e
ntal set
up
i
nvol
ved
1
tra
ns
m
itter f
or eac
h
r
ob
ot and
1
recei
ve
r
that c
ou
l
d
c
omm
un
ic
ate
con
ti
nu
ously
.
Each
r
obot
w
ould
at
te
m
pt
to
fo
ll
ow
the
t
rac
e
of
oth
e
r
r
obot
s
by
sens
i
ng
t
heir
sig
nal
stre
ng
t
hs
.
The
r
obot
c
ou
l
d
est
im
a
te
the
distance
of
nea
rb
y
r
obots
by
m
easur
in
g
the
Re
cei
ved
Si
gnal
Stren
gth
I
nd
ic
at
or
(RSSI)
of
t
he
r
ecei
ved
rad
i
o
m
essages.
Howev
e
r,
t
he
RS
SI
m
easur
e
w
a
s
ver
y
noisy
,
e
sp
eci
al
ly
in
an
indo
or
env
i
ronm
ent due
to
inter
fer
e
nc
e an
d reflect
ion
s
of t
he radi
o
si
gn
al
s.
As
sho
wn
i
n
F
igure
8,
at
ar
en
a
1x6
m
,
the
tim
e
need
ed
by
the
rob
ot
1
to
t
rav
el
a
head
wa
s
m
or
e
tha
n
oth
e
r
rob
ots.
T
he
distance
be
tween
r
obot
2
and
Ro
bot
3
was
cl
os
er
,
t
he
RSSI
value
w
as
-
72
dBm
and
-
45
d
Bm
.
Af
te
rw
a
rd,
al
l
of
the
rob
ots
m
ov
ed
towa
rd
t
he
be
nd,
the
RS
SI
value
of
R
obot
1
was
inc
r
eased
.
This
occ
ur
e
d
be
cause
of
the
aren
a
was
na
rrow
a
nd
the
obsta
cl
es
wer
e
only
the
wall
s.
If
a
high
RSS
I
value
was
obser
ved,
this
in
dicat
ed
that
the
rob
ot
was
nea
r
the
sign
al
tra
ns
m
itter
of
oth
er
r
obots
.
T
her
e
for
e,
the
rob
ot
would
be
giv
en
a
sm
aller
traveli
ng
distance.
O
nce
they
reache
d
th
e
end
of
the
trace,
they
wo
ul
d
travel
furthe
r
into
t
he
unknow
n
en
vir
on
m
ent
un
ti
l
they
cou
l
d
m
ai
ntain
the
m
i
nim
a
l
connecti
on
t
o
the
rest
of
t
he
gro
up.
T
hen,
a
gr
eat
e
r
tra
ve
l
distance
would
be
giv
e
n
t
o
the
r
obot
w
hen
the
obser
ved
RSS
I
wa
s
sm
a
ll
.
This
was d
on
e
to r
e
duce the
e
xecu
ti
on ti
m
e and r
e
duce the
po
s
sibil
it
y of
e
rror
s
close
d dis
ta
nce.
Figure
8.
RSS
I
v
al
ue
s fr
om
e
xp
e
rim
ental
en
vir
on
m
ent
4.3
.
E
xp
eri
m
ent
al R
es
ults
On F
uz
z
y
Robo
t
F
or Coor
dinat
i
on
T
he
ex
pe
rim
en
t
was
cond
ucte
d
in
a
la
borato
ry
with
an
e
nv
i
ronm
ent
con
dit
ion
e
d
in
th
ree
aren
as
1x
6
m
,
2
×
4
m
and
3x
4
m
.
Swa
rm
ro
bot
wit
h
th
ree
ide
ntic
al
r
obots
with
dif
fer
e
nt
col
ors
m
ov
ed
to
ge
ther
i
n
ta
nd
em
with
gr
eat
coord
i
nation
am
on
g
the
m
.
At
the
m
o
m
ent
the
ro
bo
t
coo
r
din
at
e
d
in
a
fr
ee
and
broa
d
env
i
ronm
ent
(
2x4
m
),
t
hen
the
r
obot
m
ov
ed
t
o
tu
r
n
ac
cordin
g
t
o
l
ogic
.
The
r
obot
w
ou
l
d
go
to
gethe
r
si
m
ultaneou
sl
y
if
in
f
ront
of
t
hem
,
there
wa
s
a
hitch
i
n
th
e
form
of
a
w
al
l,
then
the
r
obot
sl
ow
e
d
down
t
he
m
ot
ion
a
nd
the
n
tu
r
ned
m
aneu
ve
rin
g
(
hard
velocit
y).
This
pro
ve
d
that
f
uz
zy
log
ic
was
capab
le
to
w
ork
as
a
con
t
ro
ll
er
on
t
he
m
ob
il
e
ro
bot
as
it
cou
ld
pro
vid
e
a
goo
d
m
otion
respon
s
e.
I
n
Fig
ure
9,
the
m
ob
il
e
robo
t
m
ov
e
m
ent
res
pons
e
was
s
hown
as
t
he
c
ha
nge
of
le
ft
a
nd
r
igh
t
P
WM
m
oto
r
.
T
he
rob
ot
di
d
not
hit
the
wall
or
oth
e
r
r
obots
du
rin
g
the
m
ov
e in the
free e
nv
i
r
on
m
ent.
-10
0
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
RSSI
(
d
b
m)
Ti
me
Ro
bot
1
Ro
bot
2
Ro
bot
3
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
4
8
–
5
7
56
The
data
of
m
oto
r
velocit
y
vs
tim
e
per
seco
nd
ar
e
show
n
Fi
gure
9.
It
can
be
seen
that
if
t
he
sp
ee
d
of
P
W
M
was
fas
t
and
t
he
rob
ot
wen
t
str
ai
ght,
it
m
eant
that
there
was
no
obsta
cl
e;
the
rob
ot
was
i
n
a
safe
po
sit
io
n.
Wh
il
e
in
an
in
secu
r
e
adj
ace
nt
po
si
ti
on
,
the
r
obot
would
re
du
ce
t
he
spe
ed
a
nd
r
otate
the
d
irect
ion
to
fin
d
a sa
fe
posit
ion
.
(a)
(b)
(c)
Figure
9. RSS
I
value
s fr
om
the exper
im
ental
en
vir
on
m
ent (a
)
r
obot
1 (
b)
r
obot
2 (c) r
obot
3
5.
CONCL
US
I
O
N
This
pa
per
re
porte
d
the
anal
ysi
s
of
a
swarm
ro
bot
coordi
nation
usi
ng
f
uzzy
log
ic
to
con
t
ro
l
the
coor
din
at
io
n
a
m
on
g
in
div
id
uals
in
the
s
war
m
.
The
re
la
ti
on
sh
i
p
between
a
direct
ion
of
m
otion
in
th
e
coor
din
at
io
n
a
nd
exc
ha
ng
e
of
i
nfor
m
at
ion
th
rou
gh
wi
reless
c
omm
u
nicat
ion
was
est
ablished
.
I
n
this
exp
e
rim
ent,
robo
ts
c
oor
din
at
ed
to
oth
e
r
m
e
m
ber
s
us
in
g
wi
reless
c
omm
un
ic
at
ion
.
E
ach
r
obot
wou
ld
try
to
fo
ll
ow
the
othe
r
r
obots
by
se
ns
in
g
their
si
gnal
stren
gth.
T
he
r
obot
co
uld
est
i
m
at
e
the
di
sta
nce
to
the
near
est
rob
ot b
y m
easur
i
ng the
Acce
pted Si
gnal
Str
eng
t
h Indica
to
r
(
RSS
I) of t
he recei
ve
d
in
f
orm
at
ion
.
The
i
nfor
m
at
ion
receive
d
by
each
rob
ot
is
ba
sed
on
in
put
f
ro
m
the
en
vir
onm
ent,
w
hich
i
s
co
ntr
olled
us
in
g
f
uzzy
lo
gic.
I
n
this
e
xperim
ent,
it
ca
n
be
c
oncl
ude
d
that
the
siz
e
of
the
e
nvir
onm
ent
aff
ect
ed
the
coor
din
at
io
n
of
the
r
obot
m
ov
em
ent.
In
the
na
rro
w
a
ren
a
,
the
r
obot'
s
m
ov
em
ent
was
s
lowe
r
tha
n
the
la
rg
e
r
aren
a
.
This
slo
wer
s
peed
was
du
e
to
each
r
obot
lo
wer
e
d
th
ei
r
sp
eed
to
fi
nd
a
safe
posit
ion
to
av
oid
obst
acl
es.
The
s
wa
rm
robo
t
m
ov
e
d
i
n
t
he
sam
e
direct
ion
i
n
ta
nd
em
and
eac
h
r
obot
m
a
intai
ned
t
he
ir
posit
io
n
wi
thin
a
certai
n dist
anc
e.
Our
f
uture
wor
k
will
f
ocu
s
on
sw
arm
ro
bot
s
that
can
c
oor
din
at
e
in
m
aking
a
nd
ke
epi
ng
f
or
m
at
ion
.
Fu
rt
her
resear
ch
that
can
be
dev
el
op
e
d
i
s
con
t
ro
ll
in
g
f
or
m
at
ion
with
sta
ti
c
and
dynam
ic
env
iro
nm
ental
conditi
ons.
T
hus, bet
te
r res
ults
can
b
e
achie
ve
d for
furthe
r
r
esearch
.
ACKN
OWLE
DGE
MENTS
Au
t
hors
tha
nk
the
Mi
nistry
of
Re
searc
h,
Tech
no
l
og
y
a
nd
Nati
on
al
E
du
cat
io
n
(R
ISTE
K
DIKT
I
)
Ind
on
esi
a a
nd
Stat
e o
f Poly
te
chn
ic
Sr
i
wij
ay
a for
t
heir fina
ncial
sup
port in Gra
nts Pro
j
e
ct
. Th
is
pa
per
i
s one
of
our
P
h.d.
pr
oject
.
O
ur
ea
r
ne
st
gr
at
it
ude
al
s
o
go
es
to
al
l
r
esearche
rs
i
n
Tel
ecom
m
un
icati
on
a
nd
Sig
na
l
an
d
Con
tr
ol
Lab
or
at
or
y,
Ele
ct
ric
al
Eng
inee
rin
g,
Po
ly
te
ch
nic
Sr
iwijaya
w
ho
pro
vid
e
d
com
pan
io
nship
an
d
s
har
i
ng
of their
knowle
dg
e
.
REFERE
NCE
S
[1]
R.
Dori
y
a,
S.
Mi
shra,
and
S.
Gupta,
“
A
brie
f
surve
y
and
anal
y
s
is
of
m
ult
i
-
robot
c
om
m
unic
at
ion
a
nd
coor
dina
t
ion,”
Int.
Con
f. Comput.
Comm
un.
Autom
.
,
pp
.
1014
–
1
021,
2015
.
[2]
J.
C.
Barca,
Y.
A.
Seker
ci
og
lu,
J.
C.
Barc
a,
and
Y.
A.
Seker
ci
o
glu,
“
Sw
arm
ro
boti
cs
rev
i
ewe
d
Sw
arm
roboti
cs
rev
ie
wed
,
”
no.
Jul
y
2012,
2013.
[3]
K.
Sugihara
and
I.
Suzuki,
“
Distribut
ed
Mot
ion
Coordina
ti
on
o
f
Multi
ple
Mobi
le
Robots,”
Proc.
5th
IE
EE
Int
.
Symp.
Int
ell.
Co
ntrol
,
vo
l. 1, no.
1,
pp
.
138
–
143
,
1990.
[4]
B.
L
ei a
nd
H.
C
hen,
“
Sw
arm Robot
s Form
at
ion Control
B
ase
d
o
n
W
ire
le
ss
Sens
or
Network,”
pp
.
458
–
465,
2016.
[5]
A.
S.
Hand
a
y
ani,
N.
L.
Hus
ni,
S.
Nurm
ai
ni,
and
I
.
Yani
,
“
Form
at
i
on
Control
D
esign
for
Re
al
Sw
a
rm
Robot
Us
ing
Fuzz
y
Log
ic,”
i
n
Inte
rnat
ional
Con
fe
renc
e
on
El
e
ct
rica
l
Eng
in
ee
ring
and
Computer
Sc
ie
nc
e
(
ICECOS)
2017
II.
,
2017,
pp
.
77
–
82
.
[6]
N.
L.
Hus
ni,
A.
Silvi
a,
and
S.
Nurm
ai
ni,
“
New
Chal
le
ng
es
in
Air
Quali
t
y
Sensin
g
using
Roboti
c
Sensor
Network,
”
2013.
[7]
T.
Gunn
and
J.
Anderson,
“
D
y
namic
Het
ero
ge
neous
Team
Form
at
ion
for
Rob
oti
c
Urb
an
Sea
r
ch
and
Rescu
e,”
Proce
dia
-
Proc
edi
a
Comput
.
S
c
i.
,
vol
.
19
,
no
.
A
nt,
pp
.
22
–
31
,
20
13.
0
50
1
0
0
1
5
0
2
0
0
2
5
0
3
0
0
1
11
21
31
41
51
M
o
t
o
r
Sp
ee
d
T
im
e
R
o
bo
t
1
Lef
t
PWM
Rig
h
t
PWM
0
50
1
0
0
1
5
0
2
0
0
2
5
0
3
0
0
1
11
21
31
41
51
M
o
t
o
r
Sp
ee
d
T
im
e
R
o
bo
t
2
Lef
t
PWM
Rig
h
t
PWM
0
50
1
0
0
1
5
0
2
0
0
2
5
0
3
0
0
1
11
21
31
41
51
M
o
t
o
r
Sp
ee
d
T
im
e
R
o
bo
t
3
Lef
t
PWM
Rig
h
t
PWM
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
An
alysis o
n
sw
ar
m
r
obot c
oo
r
dinati
on usi
ng
fuzzy lo
gic
(
Ad
e S
il
vi
a Han
dayan
i
)
57
[8]
A.
Marjovi
and
L.
Marqu
es,
“
O
pti
m
al
spa
ti
a
l
fo
rm
at
ion
of
sw
ar
m
roboti
c
g
as
se
nsors
in
odor
pl
um
e
findi
ng,
”
p
p.
93
–
109,
2013
.
[9]
A.
Khem
ka
,
J.
Micha
e
l,
and
S.
Panic
ker
,
“
Sw
arm
Roboti
cs
-
Surveil
l
ance
And
Monitori
ng
Of
Dam
age
s
Caused
B
y
Motor
Acc
id
ent
s
,
”
no
.
9
,
pp
.
42
–
46,
2013
.
[10]
N.
L.
Hus
ni,
A.
S.
Handa
y
an
i,
S.
Nurm
ai
ni,
and
I.
Yani,
“
Coopera
t
ive
Sear
ch
i
ng
Strat
eg
y
for
Sw
arm
Robot,
”
in
Inte
rna
ti
onal
Co
nfe
renc
e
on
Elec
tric
al
Engi
n
ee
ri
ng
and
Compute
r Sc
ie
n
ce (
ICECOS)
2017
,
2017,
pp.
92
–
97
.
[11]
W
.
Li
and
W
.
Shen,
“
Sw
arm
beha
vior
cont
ro
l
of
m
obil
e
m
ulti
-
robots
with
wir
el
ess
sensor
ne
t
works
,
”
J
.
Ne
tw
.
Comput.
App
l.
,
v
ol.
34
,
no
.
4
,
pp
.
1398
–
1407,
201
1.
[12]
S.
Atana
sov,
“
An
over
vie
w
of
wire
le
ss
comm
unic
a
ti
on
t
ec
hnol
ogie
s
used
in
w
ire
l
ess
sensor
net
works
,
”
Int
.
Sci.
Conf.
eR
A
-
8
,
no.
ISS
N
-
1791
-
1133,
pp
.
11
–
18
,
20
13.
[13]
T.
Ishim
oto
and
S.
Hara
,
“
Us
e
of
RS
SI
for
m
ot
ion
cont
ro
l
of
wire
l
essl
y
n
et
wor
ked
robot
sw
ar
m
,
”
R
OSE
2008
-
IEE
E
Int
.
Work
.
Robot
.
Sensors
Env
iron.
Proc
.
,
no.
Octob
er,
pp.
92
–
97,
2008
.
[14]
B.
Tut
uko
and
S.
Nurm
ai
ni,
“
Sw
arm
Robots
C
om
m
unic
at
ion
-
b
ase
Mobile
Ad
-
Hoc
Network
(
MA
NET
),
”
no.
Augus
t,
pp.
20
–
21,
2014
.
[15]
A.
Anand,
M.
Nith
y
a
,
and
S.
Tsb,
“
Coordina
ti
o
n
of
Mobile
Robots
with
Master
-
Slave
Archi
t
ec
tu
re
for
a
Service
Applic
a
ti
on,
”
pp
.
539
–
543
,
2014
.
[16]
J.
Huirc
an
et
al
.
,
“
Zi
gBe
e
-
base
d
wire
le
ss
sens
or
net
work
loc
a
li
z
at
ion
for
cattle
m
onit
oring
in
gra
zi
ng
b,
”
n
o
.
Novem
ber
2010,
2014.
[17]
A.
Cornej
o
and
R.
Nagpa
l, “
Lon
g
-
Li
ved
Distribu
te
d
R
el
a
ti
ve
Lo
c
al
i
za
t
ion
of
Rob
ot
Sw
arms
,
”
201
3.
[18]
K.
Benki
c
,
M.
Mala
jn
er,
P.
Pla
ninsic
,
and
Z.
C
uce
j
,
“
Us
ing
RS
SI
val
ue
for
dista
nce
est
imati
on
in
wire
le
ss
sensor
net
works
base
d
on
Zi
gB
ee,”
Proc.
15
th
In
t. Conf
.
Syst. Si
gna
ls I
mage
Proc
ess.
,
pp
.
303
–
306
,
2008
.
[19]
S.
Nurm
ai
ni, “M
oti
on
Coord
ina
t
i
on
for
Sw
arm R
obots,”
pp
.
2
–
5.
[20]
N.
Agm
on,
C.
L.
Fok,
Y.
Ema
li
ah
,
P.
Stone,
C.
Juli
en,
and
S
.
Vishw
ana
th
,
“
On
coor
dination
in
pr
ac
t
ical
m
u
lt
i
-
robot
pa
trol,”
Proc.
-
IE
EE Int
.
C
onf.
Robot.
Au
to
m.
,
pp
.
650
–
656
,
2012.
[21]
P.
Mobade
rsan
y,
S.
Kh
anmoham
m
adi
,
and
S.
Ghae
m
i,
“
An
ef
fic
i
ent
fu
zzy
m
e
thod
for
p
at
h
pl
anni
ng
a
robot
i
n
complex
env
iron
m
ent
s,”
2013
21
st I
ran.
Conf
.
Elec
tr.
Eng
.
IC
EE
2013
,
vol
.
1
,
pp
.
2
–
7,
2013
.
[22]
S.
Nurm
ai
ni,
S.
Za
i
ton,
and
R.
Firnando,
“
Coopera
t
ive
Avoidance
Control
-
base
d
Inte
rv
al
Fuzz
y
Kohone
n
Networks Algorithm
in
Sim
ple
S
warm
Robots,”
v
ol.
12
,
no
.
4
,
201
4.
[23]
A.
Adria
ns
y
ah,
Y.
Gunardi
,
B
.
Bada
ruddin
,
and
E.
Ihsanto
,
“
Goal
-
see
k
ing
Beh
avi
or
-
base
d
Mo
bil
e
Robot
Us
in
g
Parti
cle
Sw
arm Fuzzy
Con
t
roller
,
”
TEL
KOMNIKA
(
Tele
communic
ati
on
Comput.
El
e
ct
ron.
Contro
l.
)
,
vol.
13,
no
.
2
,
p.
528
,
2015
.
[24]
G.
K.
Vena
y
agam
oorth
y
,
L
.
L.
Grant
,
and
S.
Doctor
,
“
Coll
e
ct
iv
e
roboti
c
sea
r
ch
using
hy
br
id
technique
s:
Fuzz
y
logi
c
and
sw
arm i
nt
el
l
ige
nc
e inspire
d
b
y
na
ture,”
Eng. A
pp
l. A
rti
f
.
Int
el
l
.
,
vol
.
22
,
no.
3
,
pp
.
431
–
4
41,
2009
.
[25]
J.
Yu,
C.
W
ang,
and
G.
Xie,
“
Coordina
ti
on
of
Multi
ple
Robot
ic
Fis
h
with
Ap
pli
c
at
ions
to
Underwat
er
Robo
t
Com
pet
it
ion
,
”
I
EE
E
Tr
ans.
Ind.
El
e
ct
ron.
,
vol
.
6
3,
no
.
2
,
pp
.
128
0
–
1288,
2016
.
[26]
B.
P.
Gerk
e
y
an
d
M.
J.
Mata
rić,
“
Sold!:
Aucti
on
m
et
hods for
m
ul
ti
robot
coor
d
ina
t
ion,
”
IE
EE
Tr
ans.
Robot
.
Au
tom.
,
vol.
18
,
no
.
5
,
pp
.
758
–
768
,
2002
.
[27]
G.
A.
Kam
inka
,
R.
Schec
h
te
r
-
gl
i
ck,
and
V.
Sado
v,
“
Us
ing
Senso
r
Morpholog
y
fo
r
Multi
robot
Form
at
ions,”
vol
.
2
4,
no.
2
,
pp
.
271
–
2
82,
2008
.
[28]
V.
Garg
and
M.
Jham
b,
“
A
Review
of
W
ireless
Sensor
Network
on
Lo
ca
l
izati
on
Techni
ques
,
”
In
t.
J.
Eng
.
Tr
ends
Technol
.
,
vo
l. 4,
no.
April
,
pp
.
10
49
–
1053,
2013
.
[29]
A.
W
ic
hm
ann,
B.
D.
Okkali
og
l
u,
and
T
.
Kork
m
az
,
“
The
in
te
g
rat
ion
of
m
obile
(tele
)
robotics
a
nd
wire
l
ess
sensor
net
works
:
A sur
ve
y
,
”
Comput.
C
omm
un.
,
vol
.
51
,
no.
Septe
m
ber
,
pp.
21
–
35
,
2014
.
[30]
P.
Corke,
R
.
Pet
erson,
and
D
.
R
us,
“
Loc
alizatio
n
and
nav
iga
t
io
n
assisted
b
y
ne
tworke
d
coop
erati
ng
sensors
and
robots,
”
In
t. J. R
ob.
R
es.
,
vol
.
24
,
no.
9,
pp.
771
–
7
86,
2005
.
[31]
M.
Schwage
r,
J.
McLur
kin,
and
D.
Rus,
“
Distribut
ed
Cover
age
Control
with
S
ensor
y
Fe
edba
c
k
for
Networked
Robots.,
”
Robo
t. Sc
i
.
S
yst.
,
no
.
Ju
ne
2014,
pp.
49
–
56,
2006
.
[32]
S.
Xue,
C.
Sun,
J.
Ze
ng,
Y.
Jin
,
and
R.
Cheng,
“
Eff
ect
of
Com
mun
ic
a
ti
on
Modes
to
Sw
arm
Robo
ti
c
Sear
ch,”
Ope
n
El
e
ct
r.
Elec
tron
.
Eng. J.
,
vol
.
8
,
no.
1
,
pp
.
240
–
2
44,
2014
.
[33]
I.
Nava
rro
and
F.
Mati
a
,
“
A
Prop
osal
of
a
Set
of
Metri
cs
for
Coll
ec
t
ive
Movem
ent
of
Robots,”
Proc.
Work.
Goo
d
Ex
p.
Me
thodol
.
Robot
.
Robot.
S
c
i.
S
yst.
,
2009
.
[34]
A.
Jac
off,
B.
W
ei
ss
,
and
E.
Mess
ina
,
“
Evol
uti
on
of
a
per
form
anc
e
m
et
ric
f
or
urba
n
sea
rch
and
resc
ue
robots
(2003),
”
Pe
rfor
m.
Me
tri
cs
Int
el
l
.
Syst
.
,
2003.
[35]
M.
Bal
l
eri
ni
et
al.
,
“
Inte
ra
ct
io
n
ruli
ng
ani
m
al
col
l
ec
t
ive
b
ehavior
depe
nds
o
n
topol
ogi
ca
l
ra
th
er
th
an
m
et
r
i
c
dista
nc
e:
Evi
d
en
ce
from
a
f
ie
ld
st
ud
y
,
”
Pnas
,
vo
l.
105,
no
.
4
,
pp
.
1
232
–
1237,
2008
.
[36]
M.
Haque
,
C.
R
en,
E
.
Bake
r
,
A.
Douglas
Kirkpa
tri
ck
,
J.
A.
Adam
s,
and
Ab,
“A
naly
s
is
of
Sw
arm
Co
m
m
unic
at
i
on
Models,
” in
European
Conf
ere
n
ce
on
Art
if
i
ci
al
I
nte
l
l
ige
n
ce
,
201
6,
no
.
Octob
er,
p
p.
29
–
36
.
[37]
H.
W
u,
S.
Qu,
D.
Xu,
and
C.
Che
n,
“
Prec
ise
local
iz
a
ti
on
and
form
at
ion
con
trol
of
sw
arm
robots
vi
a
wire
l
ess
sensor
net
works
,
”
Math
.
Probl
.
Eng.
,
vo
l.
2014
,
2014
.
[38]
R.
K.
Ramachandran
and
S.
Berman,
“
The
Ef
fec
t
of
Com
m
uni
cation
Topol
o
g
y
on
Scalar
Fi
el
d
Esti
m
ation
b
y
Networke
d
Rob
oti
c
Sw
arms
,
”
p
p.
3886
–
3893
,
2
017.
[39]
H.
Mansor,
A.
H.
Adom
,
and
N.
Abdul
Rahi
m
,
“
Deve
lopment
of
le
ade
r
and
foll
ower
strat
e
g
y
for
sw
arm
r
obot
appl
i
ca
t
ions,”
J
.
Teknol
.
,
vo
l. 77,
no.
28
,
pp
.
55
–
5
9,
201
5.
Evaluation Warning : The document was created with Spire.PDF for Python.